Relative dosimetry

A field output factor is defined as the ratio of absorbed dose to water in any non-reference field to that in a reference field at a given depth. In conventional broad beams, it is derived from a ratio of detector readings because of the practical independence of dosimetric quantities on field size. In small field dosimetry, however, such independence does not exist, notably for perturbation factors, and a field output factor will in most cases require an output correction factor to be applied to the measured detector reading ratio. It will thus in most cases be incorrect to report a ratio of readings as a field output factor, a mistake which is, unfortunately, all too often encountered in clinical practice as well as in the scientific literature [88]. Many examples have been published showing large discrepancies between the ratio of readings measured with different types of detectors for a particular beam compared with the actual ratio of absorbed

3 ‘Non-conventional’ in the nomenclature of this COP.

dose to water values [18, 24, 73]. These discrepancies are mainly field size and detector dependent and can be unacceptably high when measured with detectors that have a large volume compared to the size of the small field. The common practice of reporting ratios of detector readings as field output factors is a mistake that has led to much confusion, potentially to serious errors and, in some of the worst cases, to real accidents. For example, the use of inappropriate detectors for measuring field output factors without further corrections has been reported as the main cause of an accidental overdosage of patients for beams defined by the Brainlab m3 micro MLC [89]. A comparison of beam data measured in different centres in France with microchambers for the 0.6 cm × 0.6 cm beam of different Varian Clinac models, under identical measuring conditions (6 MV photons, micro MLC type, SSD, depth, type and orientation of the detector), showed a discrepancy of about 15% in the extreme values of the detector output ratios as uncorrected estimates for field output factors, as shown in Fig. 9. The report by the French Society of Medical Physics (SFPM), which was the result of a follow-up effort after this accident, advises using at least two different detectors for the measurement of field output factors [89].

FIG. 9. Detector output ratios as uncorrected estimates for field output factors determined in different centres in France for 6 MV photon beams and the Brainlab m3 micro MLC (SSD = 100 cm, depth = 5.0 cm) using three different detector types. The Brainlab WOI 10-26 data correspond to manufacturer guidance data. (Reproduced from Ref. [89] with the permission of the Institut de radioprotection et de sûreté nucléaire.)

Alfonso et al. [8] have emphasized the distinction between ratios of detector readings and ratios of absorbed dose to water values by explicitly including an output correction factor in the expression for the field output factor.

In large reference fields, output correction factors are required for detectors exhibiting an energy dependent response due to low energy scattered photons originating in the treatment head and in the phantom combined with the different mass energy absorption coefficient for those low energy photons. This is for example the case for silicon based devices such as unshielded diodes and metal oxide semiconductor field-effect transistors (MOSFETs), as described in Section 2.1.5, and it results in a more or less linear increase of the response with increasing field size as illustrated in Fig. 10. From these observations, the approach employed to obtain field output factors is to use an ionization chamber for field sizes down to the one where volume averaging sets in, and use a small detector (e.g. a diode, diamond, liquid ionization chamber or organic scintillator) for smaller fields. The field output factors derived from the measurements with the small detectors are renormalized at the smallest field size where the ionization chamber is used; this is referred to as the intermediate field method in this COP.

This method has sometimes been called “daisy-chaining” [90].

For determination of field output factors in small fields, another approach has been proposed that suggests using a large area parallel plane ionization chamber (LAC) in combination with radiographic or radiochromic film [73, 74].

From the signal produced by the two dimensional fluence distribution over the area of the LAC, the value of the dose–area product (DAP) can be determined.

With accurate film dosimetry at the same plane of measurement as the LAC, the field size and a two dimensional relative absorbed dose distribution can be determined. From this and the DAP value, the absorbed dose to water is derived at the region of interest. While this is an interesting area of research, there is not enough experience and information at present to provide guidance on this method.

2.3.2.2. Lateral beam profiles

The lateral beam profile is defined as the distribution of absorbed dose to water at the reference depth in the phantom, perpendicular to the beam axis and parallel to the phantom surface. The difficulties of measuring lateral beam profiles in small photon fields are associated with the dimension of the detector’s sensitive volume, defined as the geometrical dimension of the measuring volume in the scan direction, in relation with the beam penumbra size. Even for conventional broad fields, when tertiary collimation is used for reduction of their penumbra, the effect of the detector’s finite volume can lead to inaccuracies in the determination of the penumbra width. This becomes a crucial problem in

FIG. 10. Ratio of the relative readings of various detectors and the relative reading of a PTW 31010 Semiflex ionization chamber. The relative readings of all detectors were normalized at the value for 10 cm equivalent square field size. This figure illustrates the field size dependence of solid detectors for large fields and the perturbation of ionization chambers in small fields for (a) a PTW 31006 PinPoint ionization chamber, a Scanditronix Stereotactic Field Diode (SFD) and a Scanditronix Photon Field Diode (PFD), and (b) two Thomson Nielsen Si-MOSFETS of the same type (Mosfet1 and Mosfet2) and a PTW 60003 diamond detector (replotted from Ref. [24]). The full lines are linear fits to the data points for field sizes larger than 5 cm × 5 cm and the dashed lines extrapolations of those fits to smaller field sizes to illustrate the field size dependence of the diode response solely due to phantom scatter, i.e. ignoring the effect of fluence perturbations. (Reproduced from Ref. [24] with the permission of the American Association of Physicists in Medicine.)

very small fields, as the penumbra represents an important portion of the field.

Because for these small fields detector perturbation factors show a very steep dependence on field size, small errors in penumbra measurements can result in substantial dosimetric errors.

Suitable detectors to resolve the penumbra in small photon fields are tissue equivalent radiochromic film, diodes (stereotactic, shielded or unshielded and oriented parallel to central axis), diamond detectors, small air filled ionization chambers and liquid ionization chambers [12]. Even those detectors require special measures to avoid various artefacts (e.g. the readout procedure for radiochromic film needs to be well conducted). For scanning detectors, the orientation needs to be considered and effects of stem and cable irradiation taken into account. A method has been published to derive corrected penumbrae from measurements with a series of different sized detectors [91]. A paper by Francescon et al. [58] investigated the variation of the perturbation of various small detectors as a function of off-axis position. This is very helpful in advising on the type of detector to be used for profile measurements, but it is important to be aware that owing to detector-to-detector variations combined with the extreme sensitivity of these perturbations to detector dimensions, it is advised that these not be regarded as providing generic output correction factors for profile measurements.

Another approach which has been suggested is to deconvolve the lateral beam profile from the measured profile using Monte Carlo calculated detector specific kernels [92] or simply Gaussian kernels [93] based on the observation that despite the lateral fluence convolution kernels for many detectors being quite complicated, the dose convolution kernels are blurred by the lateral range of secondary electrons and the effects of the detector construction details are lost, making Gaussian kernels adequate.

An interesting observation was made by Underwood et al. [54], who suggested that, while most detectors either under-respond or over-respond on the central axis, i.e. in the measurement of field output factors, in small fields, this is actually compensated by an opposing over-response or under-response, respectively, in the profile tails, and that the integral dose measured for many detectors would be accurate without any output correction factors. Figure 11 illustrates this for three types of detectors. In IMRT, the integral dose contribution of a small field is indeed more important than the absorbed dose to water in the centre of the field itself. This approach would of course only make sense if the same detector is used to measure the field output factor and to measure the profile, while in practice often a combination of a point detector for the field output factor and radiochromic film for the profiles would be used.

FIG. 11. Lateral beam profiles of a 0.5 cm × 40 cm field (along the short axis of the rectangular field) in a 6 MV beam measured using three different detectors. These profiles are expressed in terms of absorbed dose in cGy per monitor unit (MU) by calibration of the detectors in a large field, denoted as “Calibrated detector reading”. This figure shows the under-response of the larger detectors around the dose maximum, as schematically explained in Fig. 3. The arrows on the right hand side vertical axis indicate the ratio of the dose length product (DLP) of the solid detectors (diamond, diode) to the DLP of the PinPoint ionization chamber, illustrating that for the diamond detector and the IBA unshielded diode, the DLP is the same, despite their differences on the axis of the field. (Reproduced from Ref. [54] with the permission of the American Association of Physicists in Medicine.)